Whais thmaximum possible delay that can be introduced by Netem? I have seethait was 2^32 us ( 71 minutes ), but have also heard that some updates to thLinux kernel has affected thability to introduced that long of a delay. I atrying to reach abou*182 second* delay to emulate a spaclink to Mars, buwhen I get to about *138 seconds* the delay is not seeby my ping packets. Is thera way to fix this so I can can geat least 182 second delay? or do I need to look for aolder version of Linux? I arunning Fedora 14 with Linux kernel 2.6.35.6-35.fc14.i686 Thanks, Jacob Edwards -- Jacob Edwards GraduatResearch Assistant MSU Electrical and Computer Engineering Department Simrall 330 jme147 amsstate.edu <jacob.matthew.edwards agmail.com> -------------- nexpar-------------- AHTML attachmenwas scrubbed... URL: <http://lists.linuxfoundation.org/pipermail/netem/attachments/20120305/6bcc7949/attachment.html> Froshemminger avyatta.com Mon Mar 5 19:25:39 2012 From: shemminger avyatta.co(Stephen Hemminger) Date: Mon, 5 Mar 2012 11:25:39 -0800 Subject: NeteMaximuDelay In-Reply-To: <CAJ6Y4e-P2jrz6UOYCwXwgnr2f6c2xNVovaOBd9Yb9TyN1QWwEg@xxxxxxxxxxxxxx> References: <CAJ6Y4e-P2jrz6UOYCwXwgnr2f6c2xNVovaOBd9Yb9TyN1QWwEg@xxxxxxxxxxxxxx> Message-ID: <20120305112539.2cef205f@xxxxxxxxxxxxxxxxxxxxxxxxxxx> OMon, 5 Mar 2012 12:03:11 -0600 Jacob Edwards <jme147 amsstate.edu> wrote: > Whais thmaximum possible delay that can be introduced by Netem? I have > seethait was 2^32 us ( 71 minutes ), but have also heard that some > updates to thLinux kernel has affected thability to introduced that > long of a delay. I atrying to reach abou*182 second* delay to emulate a > spaclink to Mars, buwhen I get to about *138 seconds* the delay is not > seeby my ping packets. > Is thera way to fix this so I can can geat least 182 second delay? or > do I need to look for aolder version of Linux? I arunning Fedora 14 > with Linux kernel 2.6.35.6-35.fc14.i686 > > Thanks, > Jacob Edwards > Thissues is thapsched_tdiff_t is long (32 bits). You could try a 64 bikernel if your hardwarallows, then the tdiff would b64 bits. Or juschange definition in net/pkt_sched.h from typedef long psched_tdiff_t; to typedef s64 psched_tdiff_t; Frojme147 amsstate.edu Mon Mar 5 22:17:28 2012 From: jme147 amsstate.edu (Jacob Edwards) Date: Mon, 5 Mar 2012 16:17:28 -0600 Subject: NeteMaximuDelay In-Reply-To: <20120305112539.2cef205f@xxxxxxxxxxxxxxxxxxxxxxxxxxx> References: <CAJ6Y4e-P2jrz6UOYCwXwgnr2f6c2xNVovaOBd9Yb9TyN1QWwEg@xxxxxxxxxxxxxx> <20120305112539.2cef205f@xxxxxxxxxxxxxxxxxxxxxxxxxxx> Message-ID: <CAJ6Y4e8QPpnmFYfW0POECX_OjoBo6PiCWARariz_Qom8icr80A@xxxxxxxxxxxxxx> OMon, Mar 5, 2012 a1:25 PM, Stephen Hemminger <shemminger at vyatta.com>wrote: > OMon, 5 Mar 2012 12:03:11 -0600 > Jacob Edwards <jme147 amsstate.edu> wrote: > > > Whais thmaximum possible delay that can be introduced by Netem? I > have > > seethait was 2^32 us ( 71 minutes ), but have also heard that some > > updates to thLinux kernel has affected thability to introduced that > > long of a delay. I atrying to reach abou*182 second* delay to > emulata > > spaclink to Mars, buwhen I get to about *138 seconds* the delay is > not > > seeby my ping packets. > > Is thera way to fix this so I can can geat least 182 second delay? or > > do I need to look for aolder version of Linux? I arunning Fedora 14 > > with Linux kernel 2.6.35.6-35.fc14.i686 > > > > Thanks, > > Jacob Edwards > > > > Thissues is thapsched_tdiff_t is long (32 bits). > You could try a 64 bikernel if your hardwarallows, then the tdiff would > b64 bits. Or juschange definition in net/pkt_sched.h from > > typedef long psched_tdiff_t; > to > typedef s64 psched_tdiff_t; > I havfound thpkt_sched.h on my Fedora machine, which was located in the usr/include/linux/ directory. Butheris no typedef long psched_tdiff_t; definitioin thcode. Even if I did find that and recoded it, how would I recompilisince it is built into the kernel? -------------- nexpar-------------- AHTML attachmenwas scrubbed... URL: <http://lists.linuxfoundation.org/pipermail/netem/attachments/20120305/9315088d/attachment.html> Froacmorton aatt.com Mon Mar 5 23:09:53 2012 From: acmortoaatt.com (Al Morton) Date: Mon, 05 Mar 2012 18:09:53 -0500 Subject: NeteMaximuDelay In-Reply-To: <CAJ6Y4e8QPpnmFYfW0POECX_OjoBo6PiCWARariz_Qom8icr80A@mail.g mail.com> References: <CAJ6Y4e-P2jrz6UOYCwXwgnr2f6c2xNVovaOBd9Yb9TyN1QWwEg@xxxxxxxxxxxxxx> <20120305112539.2cef205f@xxxxxxxxxxxxxxxxxxxxxxxxxxx> <CAJ6Y4e8QPpnmFYfW0POECX_OjoBo6PiCWARariz_Qom8icr80A@xxxxxxxxxxxxxx> Message-ID: <201203052308.q25N8pPL026436@xxxxxxxxxxxxxxxxxxxx> Jacob, as anlalternative: Armultiplemulators available, or extra eth ports for multiplpasses through thsame emulator? Al A05:17 PM 3/5/2012, Jacob Edwards wrote: >I havfound thpkt_sched.h on my Fedora machine, which was located >ithusr/include/linux/ directory. >Butheris no > >typedef long psched_tdiff_t; > >definitioin thcode. Even if I did find that and recoded it, how >would I recompilisince it is built into the kernel?
